Journal of Taphonomy VOLUME 4 (ISSUE 2)

PROMETHEUS PRESS/PALAEONTOLOGICAL NETWORK FOUNDATION (TERUEL) 2006 Available online at Broughton www.journaltaphonomy.com et al. Journal of Taphonomy VOLUME 4 (ISSUE 2)

The Taphonomy of Owl-Deposited Fish Remains and the Origin of the Homestead Cave Ichthyofauna

Jack M. Broughton*, Virginia I. Cannon, Shannon Arnold Department of Anthropology, University of Utah, 270 South 1400 East, RM 102, Salt Lake City, UT 84112, USA

Raymond J. Bogiatto Department of Biological Sciences, California State University, Chico, CA 95929-0515, USA

Kevin Dalton Department of Anthropology, California State University, Chico, CA 95929, USA

Journal of Taphonomy 4 (2) (2006), 69-95. Manuscript received 29 September 2005, revised manuscript accepted 18 October 2005. We report a case of fish prey dominating the diet of modern barn owls (Tyto alba) and conduct a descriptive taphonomic analysis on the fish remains the owls deposited. From a sample of 14 barn owl pellets collected on the floor of a Nevada barn, we identified 3294 tui chub (Gila bicolor) bones. These remains, derived from very small-sized fish, comprised nearly 90% of the total pellet NISP and were characterized by relatively complete skeletal part representation, and minimal bone fragmentation and digestive surface damage. We use this data-set, along with tui chub samples deposited by other agents, to evaluate the origin of fish remains derived from late Quaternary deposits of Homestead Cave, located in the northern Bonneville Basin, Utah. Quantitative comparisons of skeletal part representation and digestive damage show that the Homestead Cave fish assemblage is statistically indistinguishable from the owl-derived collection but different from chub samples originating from coyote (Canis latrans) scat and human faeces. Qualitative evaluations of other agents also suggest an owl-based origin of the fauna. Our analysis calls attention to the important role that owls can play in depositing fish in caves and rockshelters and provides useful information to researchers interested in deciphering the taphonomic history of fish remains recovered from these settings around the world. Key words: BARN OWL, FISH REMAINS, GREAT BASIN, HOMESTEAD CAVE, OWL PELLETS, RAPTOR DEPOSITS, TAPHONOMY, TUI CHUB

Introduction Muñoz & Casadevall, 1997; Rick et al., 2001; Smith, 1985) yet their value to inform Fish remains are frequently represented in on ancient human foraging behavior or cave and rockshelter faunas around the paleoenvironmental conditions in these world (e.g., Driver, 1988; Nagaoka, 2005; contexts often depends on understanding

69 Owl-deposited fish remains

Utah (Figure 1; Broughton, 2000a, b; Broughton et al., 2000; Grayson, 1998, 2000a, b; Livingston, 2000; Madsen, 2000; Madsen et al., 2001). At Homestead Cave, the lowest late Pleistocene-aged stratum produced a substantial small vertebrate fauna that included nearly 14,000 identified fish specimens. Both roosting owls and piles of owl pellets—regurgitated casts of indigestible fur and bone— were observed in the cave when the excavation began (Figures 2-3). Signs of other bone accumulating agents (people, canids) in the assemblage were limited, and there was no evidence for water-laid sediments. Finally, the deep, enclosed cave context would seem to have precluded the occupation of such piscivorous raptors as bald eagles (Haliaeetus leucocephalus) or osprey (Pandion haliaetus). Figure 1. Map of southwestern USA showing the Great All of this pointed to owls as the depositional Basin, Bonneville Basin, Great Salt Lake and sites discussed in the text (from Madsen et al. 2001). 1) agent of the Homestead fish remains, but we Homestead Cave, 2) Bitner Ranch, 3)Harney Lake, 4) could find no evidence in the literature that North Hidden Cave. American owls utilize fish in appreciable quantities and no taphonomic analyses of their depositional origin. This remains a owl-deposited fish remains were available challenge, however, since the signs of for us to evaluate this suggestion. human processing can be subtle (e.g., We fill both those gaps here by 1) Butler, 1993, 1996; Van Neer & Morales documenting a case in which fish dominated 1992; Wheeler & Jones, 1989:68) and there the diets of modern barn owls (Tyto alba; has been little research conducted on the family Tytonidae) living in northwestern taphonomy of cave-derived ichthyofaunas. Nevada, and 2) conducting a taphonomic And while owls are widely recognized for analysis on the fish materials deposited by the role they play in the deposition of small those owls. In particular, we describe a mammal and bird materials in caves (e.g., variety of osteological characteristics commonly Andrews, 1990; Bochenski & Tomek, 1997; used to address issues involving the mode or Bochenski & Nekrasov, 2001; Hockett, agent of deposition with archaeological and 1995; Lyman et al., 2003), they are rarely paleontological faunas including skeletal considered as potential contributors of fish part representation, fragmentation, size remains in these settings. structure, and digestive modifications. We We were recently compelled to such a then compare taphonomic patterns of the consideration during analyses of the small Homestead Cave fish remains with those vertebrate fauna from Homestead Cave from the barn owl sample and other bone located in the northern Bonneville Basin, accumulating agents to evaluate the

70 Broughton et al.

Figure 2. Western screech owl (Megascops kennicottii) on a perch in Homestead Cave (photograph by M.W. Shaver III).

Figure 3. View of an owl pellet cone beneath the perch (Figure 2) on the floor of Homestead Cave (photograph by M. W. Shaver III).

71 Owl-deposited fish remains

Table 1. Numbers of identified specimens (NISP) owls had also been recently observed in the and minimum numbers of individuals (MNI) per structure (P. Van Ornum, pers. comm.), but taxon for the Bitner Ranch barn owl pellets. it is unknown whether they used the site for Taxon NISP %NISP MNI %MNI nesting. The varying states of pellet Gila bicolor 3294 89.2 48 73.0 preservation suggested they had accumulated Passeriformes 62 1.68 2 3.0 over an extended, but unknown, period of time. We found no additional pellets on Sturnella neglecta 7 <1 1 1.5 revisits to the barn in July of 2003 and Sorex sp. 2 <1 1 1.5 2005. Rodentia 128 3.47 - - Cursory inspection of the pellets in the Thomomys talpoides 86 2.33 5 7.5 field revealed that fish bones were an Peromyscus sp. 31 <1 3 4.5 abundant constituent. The pellets were Microtus sp. 84 2.27 6 9.0 bagged as a unit in the field and returned to Total 3694 100 66 100 the Zooarchaeology Laboratory (Department of Anthropology, University of Utah) for examination; the analyses that follow focus depositional origin of that fauna. The data on the bones derived from the aggregated and analyses presented here should be useful assemblage of 14 pellets. In the lab, the to researchers interested in deciphering the dried pellets were dissected with forceps taphonomic history of fish remains in cave and all of the osteological materials visible and rockshelter faunas around the world. under 5x magnification were separated for identification. Our identifications are based on comparisons with Recent vertebrate A case of fish dominating the diet of specimens from the Utah Museum of modern owls: the Bitner Ranch barn owl Natural History and criteria described pellets below.

Pellet sample and context Taxonomic composition

In July of 2002, we (J.M.B., R.J.B., and Fish remains, namely tui chub (Gila K.D.) collected 14 owl pellets in various bicolor; family Cyprinidae), comprise an stages of decomposition from the floor of an overwhelming majority of the pellet-derived abandoned barn at Bitner Ranch, Washoe bones, representing 89.2% of the total County, Nevada. Bitner Ranch (1747m asl) number of identified specimens (3294 of sits at the south end of a long (10 km) 3694 total NISP) and 73% of the sample of meadow (Badger Creek Meadow) in the identified individuals (48 of 66 total MNI; southern Guano Basin. Big sage (Artemisia Table 1). No other fish species is present in tridentata) dominates the vegetation the collection. community surrounding the ranch. The We distinguished tui chub from other pellets were consistent in size with those of cyprinid species (e.g., speckled dace barn owl and retained the dried saliva film [Rhinichthys osculus], Lahontan redside or cemented appearance diagnostic of [Richardsonius egregius]) by obvious tytonid pellets (König et al., 1999:26). Barn differences in the pharyngeal tooth row

72 Broughton et al. formulae and morphology of the bony arch 1990). Smaller and larger prey —less than (see Butler, 2001; Follett, 1970; Smith, 10 and up to 500 g— are also taken 1985). Given that all of the pharyngeals depending on local densities and the present in the sample were from tui chub, availability of prey in the preferred size we assigned the remaining cyprinid range (Derting & Cranford, 1989; Ille, elements to that taxon as well. 1991; Janes & Barss, 1985; König et al., The chubs almost surely originated 1999:195; Marti, 1974:54). In the Great from Badger Creek, an intermittent stream Basin, voles (Microtus) are the favored prey located ~100 m west of the barn. By mid- in most areas with other important taxa summer, most sections of Badger Creek are including Sorex, Thomomys, Perognathus, dry, at which time fish are restricted to Dipodomys, Reithrodontomys, and Mus deeper water-retaining pools in the stream (Marti, 1990). bed. Sheldon tui chub (Gila bicolor Barn owls are known, however, to eurysoma), a subspecies endemic to the exploit unusual prey opportunities and Guano Basin of northwest Nevada and switch to new prey species as they become southeast Oregon (Williams and Bond, available. Barn owls roosting in caves, for 1981), is the only fish taxon that has been instance, will prey heavily upon bats that collected from the creek in recent surveys share their sites (Andrews, 1990:29). And (E. Flores, pers. comm.). As is typical for barn owls residing near seabird nesting tui chub populations inhabiting springs, colonies have been known to prey ponds, and small creeks, Sheldon tui chub intensively on small seabirds (e.g. Leach’s are very small with the average length of Storm-petrels [Oceanodroma leucorhoa]; adults only 63.8 and 70.6 mm, for males Bonnot, 1928). Frogs and other amphibians and females, respectively (E. Flores, pers. are occasionally taken in abundance as well comm.; Williams & Bond, 1981). (Lenton, 1984; Esteban et al., 1998). The only previously published Fish and barn owl foraging behavior detailed account of barn owls taking fish was provided by Gallup (1949:150) near To our knowledge, this case represents the Escondido, in southern California (San only well-documented example of a North Diego Co.). Gallup noted that: “Barn owls American owl species relying heavily on are common birds in this region. I have seen fish prey. And while we have no data on the as many as 30 on the beach at night feeding overall role of fish in the annual diet of on grunion when these were running”. these owls, or how and in what contexts California grunion (Leuresthes tenuis) are they were captured, it is clear that when well known for their unusual spawning they roosted or nested in the Bitner Ranch habits in which adults spawn en masse on barn, tui chub was the principal prey species sandy beaches at night during high tides used. Of course, micromammals inhabiting (Eshmeyer et al., 1983:119). The few open areas —usually rodents and insectivores available accounts of other North American ranging between 20 and 70 g— are by far the owls taking fish (e.g., great horned owl most common prey taken by barn owls over [Bubo virginianus]) also suggest that fish their nearly cosmopolitan range (Johnsgard, prey are exploited opportunistically when 2002:69; König et al., 1999:195; Marti, capture costs are substantially reduced as

73 Owl-deposited fish remains Jaws Jaws Neurocranium

* Includes both metapterygoids and mesopterygoids. Table 2. Tui chub skeletal element representa ntmcl Skeletal element Anatomical ein region oa o ein 7 37 8. 188.5 377 379 Total for region 63 Hyomandibular 63 64 Quadrate 85 31.5 64 Articular 315.3 85 32.0 68.48 Dentary 603 42.5 64 33 69.56 Premaxilla64 32 611 92.39 Total for region 32.0 16.0 3 69.57 34.78 3 3.0 6.52 Vomer 29 3 29 3 Supraorbital 1.5 14.5 Prootic 28 3.26 28 Pterosphenoid 31.52 65 20 14.0 Epiotic 65 20 Pterotic 30.43 32.5 38 10.0 38 Sphenotic 22 70.65 21.74 20 19.0 22 Prefrontal 20 11.0 41.30 Palatine 10.0 123 23.91 Post-temporal 21.74 16 123 Otolith 16 30.8 8.0 66.85 17.34 Basioccipital 42 41 Parasphenoid 29 43 41.0 22 Exoccipital 43 48 22.0 89.13 Supraoccipital 48 33 21.5 13 33 24.0 Parietal 47.83 13 46.74 16.5 Frontal 52.17 13.0 35.87 28.26 Maxilla 70 69 34.5 75.00 26 13.0 metapterygoid* 26 28.26 Mesa- Ethmoid 10 10 10.0 21.74

NISP MNE MAU NISP MNE %MAU tion from the Bitner Ranch barn owl pellets. Gadttl 24 29 053 1035.3 3239 3294 Grand total Post-cranial throat Gill and

ntmcl Skeletal element NISP MNE MAU %MAU Anatomical region oa o ein 79 79 9. 290.5 1739 1779 Total for region 78.26 29 36.0 29 4.8 Hypural 72 10.43 42.39 72 19.5 Ultimate Vertebra 1 & 2 39 vertebra 39 37 Pectoral fin spine 31 Pterygiophore 37 31 48 Basipterygium 37.0 62 1.9 48 62 4.22 9 80.43 9 24.0 241.0 31.0 4.5 Scapula 9.78 52.17 520 67.39 Coracoid 525 Supracleithrum 32 Total for region 32 16.0 34.78 37 37 49 Branchiostegal 49 18.5 Urohyal 15 24.5 40.22 Epihyal 15 ray 53.26 77 1 7.5 Hypohyal 1 77 60 1.0 16.30 Ceratohyal 60 2.17 38.5 Basihyal 10.0 83.69 18 Pharyngeal 21.74 18 61 Subopercle 9.0 92 57 Preopercle 91 41 19.57 28.5 41 Interopercle 46.0 61.96 20.5 100.00 44.56 iaevr. 24 24 28 71.30 32.8 1214 tus 92 1214 Weberian appara- timate vert. Vertebra 3 - penul- 82 41.0 Cleithrum 114 89.13 84 42.0 91.30 Opercle 74 74 37.0 80.43

74 Broughton et al. when fish are stranded—dead or Table 3. Fragmentation of selected tui chub elements alive—on shorelines or partially derived from the Bitner Ranch barn owl pellets. exposed in very shallow streams Skeletal element Degree of element completeness (Bogiatto et al., 2003; Broughton, 2000a:104-105; Errington, 1940). 0-33% 33-67% 67-100% % > 67%* In sum, Barn owls usually Basioccipital 6 1 34 83 take small mammals but are clearly Articular 0 0 63 100 opportunistic hunters and will take Opercle 16 16 26 45 a variety of prey, including fish Pharyngeal 2 9 80 88 that fall within their targeted size Ceratohyal 0 2 75 97 range. We hypothesize that the Vertebra 15 56 1111 94 Bitner Ranch barn owls scavenged Total 39 84 1389 92 tui chub carcasses that resulted *Column 3/∑ columns 1-3. from periodic summer dessications or winter freezes of Badger Creek.

The taphonomy of barn owl-deposited Skeletal part representation fish remains We measured relative skeletal abundances Previous barn owl taphonomic work with from the pellet-derived tui chubs by small mammal and avian prey shows that comparing the number of observed elements ejected pellets contain materials typically to the number expected. These values are characterized by a high degree of skeletal derived from minimal animal units completeness, with comparatively low degrees (MAU)— counts per element normed by the of bone fragmentation, and corrosion caused number of times the latter occurs in the by digestion (e.g., Andrews, 1990; Dodson body of a single individual (see Binford, & Wexler, 1979; Hoffman, 1988; Kusmer, 1984; Mollhagen et al., 1972). The MAU 1990; Lyman et al., 2003; Raczynski & values are based, in turn, on MNEs, or the Ruprecht, 1974; but see Saavedra & minimum number of individual elements Simonetti, 1998 for intraspecific variation in needed to account for a collection of barn owl bone deposits). In this section, we element fragments. In this case, the best- conduct a taphonomic analysis of the tui represented element was the pharyngeal, chub remains derived from the Bitner Ranch with an MAU of 46; the abundances of all barn owl pellets describing patterns in other elements were scaled to this value, skeletal part representation, size structure, accordingly (i.e., % MAU for the fragmentation, and surface digestion damage. basioccipital = 41/46 = 89.13%). To facilitate subsequent comparisons, we In general, element survivorship is follow the methods used in Butler’s (Butler, high with 18 different elements represented 1996; Butler & Schroeder, 1998) by 60% or greater of the expected values taphonomic work with tui chub materials (Table 2). Moreover, each major region of deposited by mammalian predators. the skeleton is represented by an element with a value greater than 89%, suggesting

75 Owl-deposited fish remains

Figure 4. Tui chub elements included in the analysis of bone digestive damage (all paired elements are from the left side). Not drawn to scale. (a) Articular, lateral; (b) basioccipital, lateral; (c) ceratohyal, medial; (d) opercle, medial; (e) pharyngeal, dorso-lateral; and abdominal vertebrae, lateral (f), and posterior (g) views (from Butler and Schroe- der, 1998). that whole fish were consumed and relatively and their affect on preservation, recovery, complete skeletons were regurgitated in pellets. and identifiability (see Butler & Schroeder, Still, some elements are underrepresented. The 1998; Darwent & Lyman, 2002; Lyman & supraorbital, vomer, basihyal, coracoid, and O’Brien 1987; Nagaoka, 2005). pterygiophore, for instance, are each In sum, these skeletal part data are represented by 10% or less of their expected consistent with previous barn owl field values. research and pellet analyses focused on Although element destruction is often mammalian prey and suggest that small fish correlated with structural bone density (e.g., prey are swallowed whole with the Lyman 1994; Butler 1993), the limited materials contained within ejected pellets available density data suggests this is not a being characterized by a high degree of factor here. Using bone density values for skeletal completeness. large sucker (Catostomus macrocheilus, family Catostomidae; from Butler 1996), a Tui chub size structure species in the same order (Cypriniformes) as tui chub, there is no correlation between We reconstructed the size profile of the barn element survivorship (%MAU) and mineral owl-deposited tui chubs by using the bone density in this assemblage (rs = 0.13, P logarithmic regression formula: = 0.64). The underrepresented elements in ln SL = 2.51260 + .90851(ln OL), the collection are, however, the smaller, where SL is the estimated standard seemingly more delicate bones of the length (end of snout to end of hypural skeleton and we suspect their scarcity is bone), and OL is the opercle length related to variables such as size and shape (maximum length of anterior border of the

76 Broughton et al.

Table 4. Frequency (NISP) of selected tui chub elements showing surface digestion damage from the Bitner Ranch barn owl pellets.

Skeletal element Pitting Rounding Deformation Present Absent % present Present Absent % present Present Absent % present Basioccipital - - - 7 34 17.1 - - - Articular 0 63 0.0 ------Opercle 0 58 0.0 ------Ceratohyal 9 68 11.7 ------Pharyngeal - - - 10 81 11.0 - - - Vertebra - - - 237 944 20.07 62 1116 5.26 Total 9 189 4.5 254 1059 19.35 62 1116 5.26 opercle bone). This formula was derived by <1/3; 1/3 to 2/3; and >2/3. Following Butler Butler (1996:704) from a modern sample of and Schroeder (1998), we selected for this tui chubs (n = 143) collected from western analysis a subset of elements that exhibit Nevada. different shapes and sizes and are located at A total of 46 opercles in the Bitner different parts of the skeleton: the Ranch tui chub sample were intact enough basioccipital, articular, opercle, pharyngeal, to measure their maximum anterior lengths. ceratohyal, and vertebra (Figure 4). A total Consistent with recent fish survey work in of 1449 specimens were examined for Badger Creek, the represented individuals fragmentation (Table 3). These data suggest are very small fish, ranging between 88.41 a minimal degree of fragmentation: 92% of and 167.67 mm SL, with a mean of 109.931 the elements examined were at least 2/3 complete. mm SL. The distribution does not differ There is some variation in completeness among significantly from a normal one (Kolmogorov- the elements, however, with, not surprisingly, Smirnov normality test; X2 = 2.78, P = 0.49). the widest, thinnest element in the sample— Using established length-weight relationships the opercle—showing the greatest degree of for tui chub (Kimsey, 1954:407), fish of fragmentation. these lengths correspond to weights between about 10 and 75 g —well within Digestive processes and bone surface the preferred prey size range of barn owl damage mammalian prey. So while the tui chub is an unusual prey type for barn owls, the size Bones and teeth ingested by avian and range of the exploited individuals is typical mammalian predators often show distinctive for the species. surface modifications caused by mastication, acid corrosion, or physical abrasion with Element fragmentation associated gritty food items. Importantly, the degree of such damage has been shown To measure element fragmentation here, we in many cases to be predator-specific (e.g., assigned selected elements to one of three Andrews, 1990; Bochenski & Tomek, 1997, categories of skeletal completeness based Mayhew, 1977). Following Butler & on the percentage of the element present: Schroeder’s (1998) work with tui chub

77 Owl-deposited fish remains

the articular, opercle and ceratohyal (Figure 5); rounding was recorded for the basioccipital, pharyngeal, and vertebrae (Figure 6); and deformation was observed on the vertebrae (see Butler & Schroeder [1998] for further details on the nature of these attributes). The entire element surfaces were examined for these features under 8-50 x magnification with a dissecting microscope. The data show that the barn owl- derived chub elements exhibit, with some exceptions, little evidence of digestion damage (Table 4). The presence of pitting and vertebral deformation —two out of the three damage categories— was observed on less than 5% of the elements examined. Rounding, however, was a bit more commonly exhibited on these materials, occurring on 19% of the elements studied. These data are consistent with previous taphonomic studies on barn owl deposited mammalian prey that have also shown low frequencies of digestive-based surface damage. This is apparently due to the fact Figure 5. (a) Unmodified tui chub ceratohyal (35 x), that barn owls swallow small prey whole lateral view, from the Bitner Ranch barn owl pellets. and produce gastric secretions with (b) Tui chub ceratohyal (35 x), medial view, from the Bitner Ranch barn owl pellets showing pronounced comparatively low acidity (Andrews, 1990; pitting. Duke et al., 1975; Kusmer, 1990; Mayhew, 1977). This generally low degree of materials derived from mammalian scat, we digestive damage contrasts sharply with selected the basioccipital, articular, opercle, mammalian predators and many other pharyngeal, ceratohyal, and vertebra from raptors (e.g., Andrews, 1990; Duke et al., the Bitner Ranch tui chub materials for 1975; Kusmer, 1990; Schmitt & Juell, analysis of bone digestion damage (Figure 1994). 4). Butler & Schroeder (1998) found that the primary digestive damage attributes, Owls and the depositional origin of the including pitting, rounding, and deformation Homestead Cave ichthyofauna — features commonly reported and described for both mammalian scat and The barn owl-derived fish data set can now avian pellet samples— were variously be employed along with cyprinid samples evident on this subset of tui chub elements. deposited by other predators to evaluate the Accordingly, pitting was examined here for origin of the Homestead Cave ichthyofauna.

78 Broughton et al.

Homestead Cave is a limestone cavern located on the northwestern margin of the Lakeside Mountains, several kilometers west of Great Salt Lake (Figs. 1-3; 7) in the northern Bonneville Basin. The cave sits at an elevation of 1406 m and is 17 m in length and varies from 1 to 4.5 m wide and 5 to 6 m high. No active springs or permanent streams are located near the site. In 1993 and 1994, a 1 m2 sample column was excavated to a depth of ~3 m toward the back of the cave. Materials excavated from the column were passed through a nested series of 1/4” (6.4 mm) and 1/8” (3.2 mm) screens (Madsen, 2000). Analyses of the site’s massive vertebrate faunal data set have substantially increased our understanding of late Quaternary paleoecology in the Great Basin (e.g., Broughton, 2000a, b; Grayson, 1998, 2000a, b; Livingston, 2000; Madsen, 2000; Madsen et al., 2001). The fish remains from Homestead Cave have played an important role in those analyses as the site provided the largest, Figure 6. (a) Unmodified tui chub abdominal vertebra well-stratified, and richest late Pleistocene (79 x), anterior view, from the Bitner Ranch barn owl pellets. (b) Abdominal vertebra (79 x), anterior view, and Holocene ichthyofaunal sequence from from the Bitner Ranch barn owl pellets showing pro- the entire Great Basin (Broughton 2000a, b; nounced rounding. Note rounded, scalloped edge on Broughton et al., 2000). Perhaps of broadest the right. interest, the fish-based trends in moisture history have been shown to co-vary are stranded dead or dying on shorelines— positively with paleontologically derived spikes in fish abundances across the fluctuations in artiodactyl populations and Homestead Cave deposits should reflect ultimately patterns in human foraging periods characterized by enhanced scavenging behavior and associated hunting technology opportunities. Since this is a closed-basin in the region (Byers & Broughton, 2004; lake context, these would result from Byers et al., 2005). climate-based recessions from high-stands Importantly, the paleoclimatic inferences of Lake Bonneville or Great Salt Lake and derived from the Homestead ichthyofauna are the ensuing fish die-offs associated with based on the suggestion that owls deposited increases in lake-water salinity and the fish in the cave. Specifically, it is argued temperature. Peaks in fish abundances in the that insofar as owls only utilize fish in cave deposits would not, however, be tied to unusual circumstances —such as when they recessions from high-stands if the fish were

79 Owl-deposited fish remains collected and deposited by active piscivores lagomorphs; and 78% of these specimens capable of extracting live fish from open represent taxa smaller than Lepus (Grayson, water (e.g., humans, osprey, bald eagle). 2000a). Below, we evaluate to what degree the taphonomic patterns in the Homestead Cave Cyprinid deposits produced by known ichthyofauna are consistent with those agents derived from the barn owl-deposited fish bones described above and distinct from In addition to the barn owl collection those of other bone accumulating agents. described above, quantitative taphonomic data on cyprinid remains deposited by The Homestead Cave ichthyofauna known agents are limited to skeletal part representation and digestive damage Although the entire Homestead Cave attributes for tui chub materials derived deposit is rich in bone, fish remains were from coyote scats, modern human faeces, most abundant in stratum I, the basal and archaeological human coprolites (Butler stratum of the deposit. This layer consists & Schroeder, 1998). The coyote sample primarily of degraded organic material, consists of four scats collected along the including vast quantities of unfossilized shore of Harney Lake, southeastern Oregon small vertebrate materials. The analyses (Figure 1), after a series of tui chub die-offs. below focus on the fish remains recovered The remains from the modern human faeces from the 1/4” and 1/8” fractions of stratum resulted from taphonomically-oriented I. Six radiocarbon dates constrain the experimental work, and the coprolites deposition of the materials from stratum I include ten specimens recovered from between ~11,200 and ~10,100 14C yrs B.P. Hidden Cave, western Nevada (Figure 1). These dates correspond with the Gilbert The tui chub remains from these collections level of Lake Bonneville: the shoreline of were recovered by sieving through the lake at that time would have been 1/8” (3.2mm) and 1/16” (1.6 mm) mesh located less than ~1 km from the mouth of screens (see Butler & Schroeder [1998] for the cave (Madsen, 2000). Nine species are represented by the 13,536 identified fish specimens from stratum I of Homestead Cave (Broughton 2000a, b). Collectively, it is a deep-water lacustrine fauna, strikingly similar to that of modern Bear Lake, a cold, deep, high- elevation lake located on the Utah-Idaho border. In addition to the fishes, 23 species of small mammals (Grayson, 2000; Madsen, 2000) and 28 species of birds (Livingston, 2000) have also been identified from the stratum I deposits. Over 99% of the ~184,000 identified mammalian specimens from the site are small rodents and Figure 7. View of the mouth of Homestead Cave.

80 Broughton et al.

Table 5. Element abundances by major anatomical region for tui chubs derived from different agents. Human** Human** Anatomical Barn owl pellets* Coyote** faeces (experimental) coprolites (Hidden Cave) region MNE %MAU† MNE %MAU† MNE %MAU† MNE %MAU† Neurocranium 603 36.06 349 37.47 7 6.85 151 30.03 Jaws 377 68.29 219 62.83 6 16.60 84 47.28 Gill and Throat 520 47.63 168 51.67 7 22.18 111 56.07 Post-cranial 1679 52.63 1017 61.50 61 45.99 433 46.38 Total 3179 1753 81 779 * This study. ** Data from Butler & Schroeder (1998). † %MAU is the mean of %MAU values derived for the individual elements that comprise a major anatomical region.

Table 6. Matrix of chi-square values for comparisons of tui chub skeletal part representation between different known agents (raw data are MNE values from Table 5).

Barn owl Coyote Human faeces pellets scat (experimental) Barn owl pellets – – – Coyote scat ***43.65 – – Human faeces (experimental) **16.23 *10.71 – Human coprolites (Hidden Cave) 3.32 *12.6 *12.06 * Significant at P < 0.01. ** Significant at P < 0.001. *** Significant at P < 0.000.

81 Owl-deposited fish remains Table 7. Adjusted residuals for comparisons of skeletal part representation among known-agent control samples. a. Barn owl vs. coyote scat Barn owl Coyote scat Neurocranium -0.80 0.80 Jaws -0.65 0.65 Gill and throat 6.57 -6.57 Post-cranial elements -3.51 3.51 b. Barn owl vs. human faeces (experimental) Barn owl Human faeces (experimental) Neurocranium 2.35 -2.35 Jaws 1.23 -1.23 Gill and throat 1.86 -1.86 Post-cranial elements -4.01 4.01 c. Coyote scat vs. human faeces (experimental) Coyote scat Human faeces (experimental) Neurocranium 2.51 -2.51 Jaws 1.36 -1.36 Gill and throat 0.28 -0.28 Post-cranial elements -3.09 3.09 d. Coyote scat vs. human coprolites (Hidden Cave) Coyote Human coprolites (Hidden Cave) Neurocranium 0.31 -0.31 Jaws 1.22 -1.22 Gill and throat -3.46 3.46 Post-cranial elements 1.14 -1.14 e. Human faeces (experimental) vs. human coprolites (Hidden Cave) Human faeces Human coprolites (experimental) (Hidden Cave) Neurocranium -2.38 2.38 Jaws -0.95 0.95 Gill and throat -1.40 1.40 Post-cranial elements 3.42 -3.42

82 Broughton et al. details); our quantitative comparisons with Adjusted residuals are read as standard the Homestead fishes are made with these normal deviates—values that exceed +/- collections and the barn owl sample 1.96 and +/-2.57 are significant at the 0.05 described above. Although we have reasons and 0.01 levels, respectively (Everitt, to doubt that people or coyotes played a 1992:47). The analysis shows that the barn significant role in depositing the Homestead owl and Hidden Cave collections are fauna, as we discuss below, these consistently overrepresented by gill and collections provide an important frame of throat elements compared to the coyote reference to compare with our barn owl- control (Table 7). In addition, the barn owl derived control sample. collection is underrepresented by post- cranial elements compared to the coyote Skeletal part representation sample. These patterns may be related in part to a more thorough destruction of fish Before turning to the Homestead fauna, we skeletons in the coyote digestive system, compare element abundances among the tui insofar as post-cranial elements such as chub known-agent control samples (Table vertebrae are more resistant to breakdown 5). To facilitate comparisons, we conduct from mechanical and biochemical processes. the analysis on aggregated sets of elements Vertebrae, we observe, are among the more that comprise four major regions of the fish dense elements of the cypriniform skeleton, skeleton: the neurocranium, jaws, gill and to judge from analyses of mineral bone throat, and the post-cranial skeleton. density on largescale sucker (Butler, 1996). Chi-square comparisons show that Indeed, density and survivorship (%MAU) skeletal part abundances are dramatically are uncorrelated in each of the collections distinct among these agents with one with the exception of the coyote scat sample notable exception; the barn owl and Hidden (Table 8). This suggests that the Cave coprolite samples are indistinct (Table distinctiveness of the coyote and barn owl 6). To explore the specific sources samples is due to a greater degree of density responsible for the differences in element mediated destruction in the former and is abundances between these collections, we consistent with previous taphonomic work examine adjusted residuals. These indicate with small mammal, bird, and fish prey which cells — in this case, different recovered from these predators (e.g., anatomical regions— are responsible for the Dodson & Wexlar, 1979; Andrews 1990; significant overall chi-square values. Schmitt & Juell, 1994; Butler & Schroeder,

Table 8. Spearman's rank order correlation coefficients for tui chub element survivorship (%MAU) and bone density from different agents; density values are from Butler (1996).

Assemblage Rho P Barn owl pellets 0,13 0,64 Coyote scat 0,66 0,02 Human faeces (experimental) 0,30 0,35 Human coprolites (Hidden Cave) 0,05 0,86

83 Owl-deposited fish remains

Table 9. Skeletal part abundances for the Homestead Cave (Stratum I) Cypriniformes.

Anatomical Skeletal element NISP *Normed NISP %Normed NISP region Neurocranium Ethmoid 31 31.0 20.26 Frontal 92 46.0 30.07 Parietal 81 40.5 26.47 Supraoccipital 27 27.0 17.65 Exooccipital 4 2.0 1.31 Parasphenoid 42 42.0 27.45 Basioccipital 69 69.0 45.10 Meso/meta-pterygoid 33 16.5 10.78 Palatine 94 47.0 30.72 Posttemporal 10 5.0 3.27 Sphenotic 21 10.5 6.86 Pterotic 79 39.5 25.82 Epiotic 38 19.0 12.42 Supraorbital 3 1.5 0.98 Circumorbital 3 1.5 0.98 Vomer 33 33.0 21.57 Total for region 660 431.0

Jaws Maxilla 115 57.5 37.58 Premaxilla 19 9.5 6.21 Dentary 122 61.0 39.87 Articular 51 25.5 16.67 Quadrate 58 29.0 18.95 Hyomandibular 66 33.0 21.57 Total for region 431 215.5

Gill and throat Opercle 104 52.0 33.99 Preopercle 28 14.0 9.15 Interopercle 6 3.0 1.96 Subopercle 4 2.0 1.31 Pharyngeal 306 153.0 100.00 Epihyal 28 14.0 9.15 Ceratohyal 66 33.0 21.57 Urohyal 38 38.0 24.84

84 Broughton et al.

Table 9 (continuation)

Anatomical Skeletal element NISP *Normed NISP %Normed NISP region Basihyal 9 9.0 5.88 Hypohyal 21 10.5 6.86 Branchiostegal 13 2.2 1.41 Epibranchial 1 0.2 0.10 Total for region 624 330.8

Post-cranial Cleithrum 88 44.0 28.76 Supracleithrum 3 1.5 0.98 Post-cleithrum 18 9.0 5.88 Coracoid 12 6.0 3.92 Scapula 27 13.5 8.82 Pectoral fin spine 41 20.5 13.40 Basipterygium 79 39.5 25.82 Pterygiophore 1 0.1 0.04 Vertebra 1-2 86 43.0 28.10 Other vertebra 1712 46.3 30.24 Weberian apparatus 32 32.0 20.92 Total for region 2099 255.3 Grand total 3814 1232.6

* Normed NISP represents skeletal part NISP values divided by the number of times the relevant part occurs in the tui chub skeleton (see Grayson & Frey 2004).

Table 10. Chi-square values for comparisons of tui chub skeletal part representation between different agents and the Homestead Cave Cypriniformes. Barn owl Coyote Human faeces Human coprolites pellets scat (experimental) (Hidden Cave)

Homestead Cave 4.70 **46.64 *13.34 3.58

*Significant at P < 0.01.

**Significant at P < 0.000.

85 Owl-deposited fish remains

1998). Most importantly, it also suggests show that Homestead is significantly that skeletal part representation may be used different from the coyote and the modern as a means of distinguishing fish remains human faecal samples, but indistinct from deposited by these two agents. However, both the barn owl and Hidden Cave because the barn owl and one of the human coprolite collections (Table 10). Element faecal samples were also indistinct, skeletal abundances in the Homestead collection, we part representation alone may not allow us also note, are uncorrelated with density (rs = to distinguish between owl and human- 0.25, P = 0.37) suggesting that density derived deposits. mediated attrition is not a primary factor The Homestead Cave materials used affecting these data. In addition, adjusted in this analysis include all specimens residuals show that both the coyote and identified to or below the order modern human faeces collections are Cypriniformes (Table 9). This order is significantly overrepresented by the more represented at Homestead primarilly by two durable post-cranial elements (e.g., species: one catostomid —Utah sucker vertebrae) than the Homestead sample (Table (Catostomus ardens)— and one cyprinid — 11). Hence, the Homestead assemblage Utah chub (Gila atraria). Conducting the differs from the coyote sample in much the analysis at this level was necessary because same way that the barn owl sample differs the vertebrae of cypriniform taxa are from it. notoriously difficult to identify below the Whatever the underlying causes, our order level and no attempt to do so was analyses of skeletal part representation made for the Homestead materials. An suggests that the Homestead collection is analysis focusing strictly on Utah chub, or fully consistent with a derivation from barn even cyprinid specimens, would thus not owls, and clearly inconsistent with an origin include this important component of the fish from coyote scat. An origin from human skeleton. Finally, anatomical part representation faeces remains ambiguous since one of the for the Homestead collection is based on human control samples—the modern human skeletal part NISP values; MNE data were faeces— showed a significant difference not provided in the original analysis of the with the Homestead assemblage, yet the Homestead ichthyofauna (Broughton 2000a) other —the Hidden Cave coprolites— did that we draw from here. However, Grayson not. & Frey (2004) have documented a close correspondence between NISP and MNE- Digestive processes and bone surface based measures of skeletal part representation damage in faunas from a variety of distinct contexts. We add to this here by observing that the Digestion-based surface damage attributes relationship between NISP and MNE values — including pitting, rounding, and for our barnowl-derived tui chub assemblage deformation— appear to be less frequently (Table 2) is astonishingly tight (r = + 1.0, P represented in the tui chub materials derived < 0.001). from barn owl pellets, relative to the other Chi-square comparisons of element known-agent controls (Table 12). Indeed, abundances between the Homestead Cave chi-square analysis of the total number of cypriniformes and the known-agent controls damaged versus undamaged specimens

86 Broughton et al.

Table 11. Adjusted residuals for comparisons of skeletal part representation between the Homestead Cave Cypriniformes and the coyote and human faeces (experimental) tui chub samples.

a. Homestead Cave vs. coyote scat Homestead Cave Coyote scat

Neurocranium -2.34 2.34 Jaws -1.29 1.29 Gill and Throat 6.72 -6.72 Post-cranial elements -2.08 2.08

b. Homestead Cave vs. human faeces (experimental) Homestead Cave Human faeces (experimental) Neurocranium 2.04 -2.05 Jaws 1.10 -1.10 Gill and Throat 1.87 -1.87 Post-cranial elements -3.63 3.63

Table 12. Proportions of specimens exhibiting digestive damage from known-agent control samples. NISP pitted/total % pitted NISP rounded/ % rounded NISP deformed/ % deformed NISP total NISP total NISP

Barn owl 9 / 198 4.5 254 / 1313 19.3 62 / 1178 5.3

Coyote 68 / 103 66.0 135 / 538 25.0 81 / 445 18.2

Human faeces 3 / 5 60.0 1 / 23 4.3 14 / 19 73.6 (experimental) Human coprolites 35 / 45 77.8 68 / 196 34.7 3 / 154 1.9 (Hidden C.) Homestead Cave 7 / 101 6.9 287 / 1764 16.3 88 / 1539 5.7

87 Owl-deposited fish remains

Table 13. Chi-square values for comparisons of the numbers of specimens damaged and undamaged by digestive processes for different known-agent control samples. Barn owl Coyote Human faeces pellets scat (experimental) Barn owl pellets – – – Coyote scat *113.10 – – Human faeces *28.94 3.40 – (experimental) Human coprolites (Hidden *62.32 0.07 2.73 Cave) *significant at P < 0.000

Table 14. Chi-square values for comparisons of the numbers of specimens damaged and undamaged by digestive processes for different known-agent control samples and the Homestead Cave Cypriniformes. Homestead Cave Barn owl pellets 1.09 Coyote scat *145.25 Human faeces (experimental) *33.17 Human coprolites (Hidden Cave) *77.07 *Significant at P < 0.000.

Table 15. Adjusted residuals for comparisons of the numbers of specimens damaged and undamaged by digestive processes for known-agent control samples and the Homestead Cave Cypriniformes. *Homestead Cave

Coyote scat -12.05 Human faeces (experimental) -5.76 Human coprolites (Hidden Cave) -8.78 *The values represented are the Homestead Cave "number damaged" cells; negative values thus indicate the underrepresentation of damaged specimens at Homestead.

88 Broughton et al. reveals that the barn owl sample is under- damage between the owl-derived tui chub represented by modified specimens compared assemblage and the Homestead Cave to each of the known-agent controls (Table cypriniform fauna. However, significant 13). Moreover, no significant differences can differences were found in these variables be detected in the frequencies of damaged between the Homestead assemblage and the specimens among the coyote and human human and coyote scat control samples. samples. This suggests that the extent of However, for reasons unclear, skeletal part digestion-based surface damage is a representation from one of the two human variable that can readily distinguish fecal samples—the Hidden Cave coprolite between deposits derived from barn owl collection— was not detectably different pellets and the scats from mammalian from either the barn owl control or the predators. Homestead fishes. Yet, overall, an The Homestead Cave cypriniformes anthropogenic origin is not supported by assemblage exhibits a generally low degree these analyses since skeletal part data from of digestion damage; pitting, rounding, and the modern human faeces was different than deformation are evident on only, 6.9, 16.3, the Homestead collection, as was the degree and 5.7% of the examined specimens, of digestive damage evident in both the respectively (Table 12). This low incidence human samples compared to Homestead. of surface damage is clearly not due to our There is no suggestion from any of these inability to observe these damage features analyses that coyotes played a role in the as a result of post-depositional weathering; deposition of the Homestead fish fauna. the remains are exceptionally well- preserved and unabraded (Figure 8; see Qualitative comparisons with other Bochenski & Tomek, 1997 for differences potential depositional processes between owl digestion damage and sediment erosion). Chi-square analysis of The analyses above were necessarily the raw numbers of specimens that do and restricted to skeletal part representation and do not exhibit these damage features shows, in fact, that the Homestead collection is indistinguishable from the barn owl sample but significantly different from each of the other comparative samples (Table 14). And adjusted residuals clearly show that the numbers of specimens modified by digestive processes are underrepresented at Homestead Cave compared to each of the mammalian control samples (Table 15).

Summary and discussion of quantitative comparisons

Figure 8. Cypriniformes caudal vertebra (79 x), No significant differences could be detected posterior view, from Homestead Cave (stratum I). This in skeletal part representation or digestive specimen shows no signs of digestive damage.

89 Owl-deposited fish remains digestion damage from the few well- such remains would thus not reflect the documented cyprinid assemblages produced substantial digestive damage that characterize by known agents. Here, we briefly evaluate these taxa. Other evidence bearing on with more qualitative criteria other potential carnivore activity from the Homestead processes for the deposition of the fauna suggests they were not a factor. For Homestead Cave fishes. These include a instance, evidence of carnivore involvement consideration of non-scatological anthropogenic is limited to three —out of over 204,000 fish deposits and other mammalian and vertebrate NISP— gnawed artiodactyl avian predators. phalanges (Broughton, 2000a; Grayson, 2000a; Livingston, 2000). Non-fecal anthropogenic deposits. Our quantitative comparison with human- Other avian agents. As a result of derived chub remains focused on materials differences in feeding behavior and collected from human faecal deposits. Of digestive physiology, pellets cast by diurnal course, human foragers routinely transport, raptors (Falconiformes) contain process, and consume fish carcasses without substantially less bone and the surviving ingesting their bones. An anthropogenic materials are characterized by a relatively origin of the Homestead fish of any sort is, high degree of fragmentation, incomplete however, very unlikely. Only a handful of skeletal representation, and pronounced cultural artifacts were found from the site surface digestive corrosion compared to and all of those were derived from late owls. These patterns, derived from analyses Holocene deposits (Shaver, 2000). Further, of avian and mammalian prey remains, are not a single burned or butchered bone was due not only to the lower gastric pH and recorded among over 204,000 identified greater bone digestion capabilities that vertebrate specimens from the site (Schmitt, characterize falconiform taxa, but to the fact 2000). that most diurnal raptors remove the flesh from larger prey and thus ingest less bone to Other mammalian predators. Taphonomic begin with (Andrews, 1990; Bochenski et work with small mammal and bird prey al., 1998, 1999; Duke et al., 1975; recovered from scats produced by a variety Hoffman, 1988; Mayhew, 1977). As a of carnivores (e.g., other canids, mustelids, result, substantial bone deposits produced ursids, felids) suggests that such deposits by Falconiformes are more likely to result are characterized by incomplete skeletal from uneaten prey remains than they are part representation, high degrees of from pellets (Bochenski et al., 1997, 1999; fragmentation, and substantial digestive Bochenski & Tornberg, 2003). surface damage (e.g., Andrews, 1990; Taphonomic work with fish remains Andrews & Evans, 1983; Schmitt & Juell, deposited by hawks and eagles is very 1994). These patterns contrast with the limited but consistent with observations minimal damage and skeletal completeness derived from bird and mammal prey of the Homestead fish and suggest a remains. And although we are unaware of scatological origin for the fauna is unlikely. work conducted on fish remains from such Carnivores can, however, transport and piscivorous North American falconiform deposit bones without digesting them and taxa as osprey or bald eagle, Stewart et al.

90 Broughton et al.

(1999) have provided detailed analyses of number of the cutthroat trout (Oncorhynchus fish materials collected from the roosts of clarki) vertebrae from Homestead exceed African fish eagles (Haliaeetus vocifer) in Kenya. 9.0 mm in maximum width, suggesting live Most notably, the fish materials —dominated by weights over 2.6 kg; some exceed 11.0 mm, cichlids and catfish (Clarias)— are characterized indicating weights over 6.2 kg (see Follet, by extensive breakage and extremely low 1980:115, 1982: figure 3). The abundance element survivorship with, overall, only of both ducks (e.g., Aythya) and mature 6.7% of the bones surviving eagle hares (Lepus) in the stratum I fauna also consumption and digestive processes. suggests the work of a more powerful The overall taxonomic composition predator. The great horned owl, well-known of the Homestead vertebrate fauna and the to roost and nest in caves, is most likely in cave microhabitat context are also this Great Basin context. This opportunistic inconsistent with diurnal raptors and other feeder has the broadest diet of any North potential avian agents. For instance, the fact American owl and takes prey ranging in that stratum I contains not only fish but an size from scorpions (Scorpiones) and mice, enormous stratigraphically intermixed to hares and porcupines (Erithizon assemblage of small, nocturnal, rodents (e.g., dorsatum; 3.8 – 18 kg); it is also known to Peromyscus, Dipodomys, Perognathus, scavenge fish (e.g., Bogiatto et al., 2003; Neotoma; Grayson, 2000a) would seem to Errington et al., 1940; Houston et al., 1998). rule out not only obligate piscivores such as Great horned owls bolt smaller prey but osprey or cormorants (Phalacrocorax), but dismember larger ones at kill sites, where also more generalist raptors such as bald heads and feet are often discarded (Houston eagle and golden eagle (Aquila chrysaetos) et al., 1998). Taphonomic work with great that focus on much larger prey items horned owl mammalian prey shows patterns (Kochert et al., 2002; Poole et al., 2002). similar to barn owls but with somewhat Finally, the deep enclosed cave context is more extensive bone damage and less inconsistent with the behavior of a number complete skeletal part representation of hawks and eagles that favor more (Dodson & Wexler, 1979; Hoffman, 1988; exposed nest site contexts, such as trees, or Kusmer, 1990). We note too that the large ledges on cliff faces, that provide size of some of the Homestead fish bones unobstructed access to nests from above and also eliminates the possibility that that they views of the surrounding landscape (e.g., entered the cave in the stomachs of fish Bechard & Schmutz, 1995; Buehler, 2000; eating birds (e.g., Phalacrocorax, Aechmophorus) England et al., 1997; Kochert et al., 2002; who had fallen prey to owls. Finally, we Poole et al., 2002; Preston & Beane, 1993). cannot exclude the possibility that other owl species contributed fish materials at Other owls. While a variety of Homestead—western screech owls (Megascops indications point to owls as the source for kennicottii), for instance, were observed the Homestead fauna, many prey in the roosting in the cave as the site was being deposit are beyond the ~500 g limit for barn excavated (Figure 2). owl and their presence suggests that other accumulators were involved. For example, a

91 Owl-deposited fish remains

Conclusions of cave-derived ichthyofaunas that exhibit these characteristics. The ability to determine the depositional Since work began at Homestead Cave, origin of fish remains recovered from cave owls have been seriously considered as and rockshelter faunas can be critical to agents responsible for the site’s rich fish understanding issues relating to ancient record and far-reaching paleoenvironmental human foraging behavior and the implications have been derived from that reconstruction of past environments. And suggestion. Those implications are supported although owls are widely recognized for the here by the documentation that fish can role they play in depositing small mammal dominate the diets of barn owls and by our and bird materials in caves, they are rarely taphonomic analysis which shows that the considered as potential contributors of fish Homestead fish sample is, overall, more remains in these settings. Our analysis of consistent with a deposition by owls than by modern barn owl pellets from Nevada any other agent. Further actualistic work documents not only that this species —a with fish remains deposited by other known widely regarded micro-mammal specialist— agents will allow still stronger conclusions will utilize fish resources but, in certain regarding the depositional origin of cave- circumstances, fish can overwhelmingly derived ichthyofaunas. dominate their diets. In this pellet sample, tui chub bones comprise nearly 90% of the total NISP. Although we lack direct Acknowledgements observations on the context of their acquisition, we suspect that the barn owls We thank P. Moyle, J. Williams, E. Flores, scavenged tui chub carcasses that resulted P. Van Ornum, E. Miskow, N. Martinez for from periodic summer dessications or helpful assistance and advice; S. Miller and winter freezes of a nearby intermittent B. Bowman for photographic assistance; stream system. and F. Bayham, V. Butler, M. Cannon, and In addition to documenting a case in L. Lyman for invaluable comments on an which fish prey dominate owl diets, we earlier draft of the manuscript. have provided the first taphonomic analysis of owl-deposited fish remains. Our analysis of 3294 pellet-derived tui chub bones References revealed a normal distribution of small- sized fish (10-75 gm), relatively complete Andrews, P. (1990). Owls, Caves and Fossils. skeletal part representation, and minimal Chicago:University of Chicago Press. Andrews, P. & Evans, E.N. (1983). Small mammal bone fragmentation and digestive surface bone accumulations produced by mammalian damage. These characteristics are similar to carnivores. Paleobiology, 9:289-307. those reported for barn owl mammalian Bechard, M.J. & Schmutz, J.K. (1995). Ferruginous prey and suggest the patterns described here Hawk (Buteo regalis). In (A. Poole & F. Gill, Eds.) The Birds of North America, No. 172. The are typical for fish remains deposited by this Academy of Natural Sciences, Philadelphia, and species as well. Owls should thus be given The American Ornithologists’ Union, Washington, serious consideration as potential contributors D.C.

92 Broughton et al.

Binford, L.R. (1984). Faunal Remains from Klasies Butler, V.L. (1993). Natural versus cultural salmonid River Mouth. New York: Academic Press. remains: origin of The Dalles Roadcut bones, Bochenski, Z.M., Huhtala, K., Sulkava, S. & Columbia River, Oregon, U.S.A. Journal of Tornberg, R. (1999). Fragmentation and Archaeological Science, 20: 1-24. preservation of bird bones in food remains of the Butler, V.L. (1996). Tui chub taphonomy and the golden eagle Aquila chrysaetos. Archaeofauna, 8: importance of marsh resources in the western 31-39. Great Basin of North America. American Bochenski, Z.M., Jussila, P., Pulliainen, E., Tornberg, Antiquity, 61: 699-717. R. & Tunkkari, P. (1998). Damage to bird bones in Butler, V.L. (2001). Fish faunal remains. In (R.L. Kelly, pellets of gyrfalcon Falco rusticolus. Journal of Ed.) Archaeological Survey and Excavations in the Archaeological Science, 25: 425-433. Carson Desert and Stillwater Mountains, Nevada. Bochenski, Z.M., Korovin, V., Nekrasov, A.E. & University of Utah Anthropological Papers, No. 123, Tomek, T. (1997). Fragmentation of bird bones in Salt Lake City, pp. 271-280. food remains of imperial eagles (Aquila heliaca). Butler, V.L. & Schroeder, R.A. (1998). Do digestive International Journal of Osteoarchaeology, 7: processes leave diagnostic traces on fish bones? 165-171. Journal of Archaeological Science, 25: 957-971. Bochenski, Z.M. & Nekrasov, A.E. (2001). The Byers, D.A. & Broughton, J.M. (2004). Holocene taphonomy of Sub-Atlantic bird remains from environmental change, artiodactyl abundances, and Bazhukovo III, Ural Mountains, Russia. Acta human hunting strategies in the Great Basin. zoological cracoviensa, 44: 93-106. American Antiquity, 69: 235-256. Bochenski, Z.M. & Tomek, T. (1997). Preservation of Byers, D.A., Smith, C. & Broughton, J.M. (2005). bird bones: erosion versus digestion by owls. Holocene artiodactyl population histories and large International Journal of Osteoarchaeology, 7: game hunting in the Wyoming Basin, U.S.A. 372-387. Journal of Archaeological Science, 32: 125-142. Bochenski, Z.M., & Tornberg, R. (2003). Darwent, C.M. & Lyman, R.L. (2002). Detecting the Fragmentation and preservation of bird bones in postburial fragmentation of carpals, tarsals, and uneaten food remains of the gyrfalcon Falco phalanges. In (W. D. Haglund & M. H. Sorg, Eds.) rusticolus. Journal of Archaeological Science 30: Advances in ForensicTaphonomy. Boca Raton, FL: 1665-1671. CRC Press, pp. 356-377. Bogiatto, R.J., Sardella, B.A. & Essex, J.J. (2003). Derting, T.L. & Cranford, J.A. (1989). Physical and Food habits of great horned owls in northeastern behavioral correlates of prey vulnerability to barn California with notes on seasonal diet shifts. owl (Tyto alba) predation. American Midland Western North American Naturalist, 63: 258-263. Naturalist, 121: 11-20. Bonnot, P. (1928). An outlaw barn owl. The Condor, Dodson, P. & Wexlar, D. (1979). Taphonomic investigations 30: 320. of owl pellets. Paleobiology, 5: 292-296. Broughton, J.M. (2000a). The Homestead Cave Driver, J.C. (1988). Late Pleistocene and Holocene ichthyofauna. In (D.B. Madsen, Ed.) Late vertebrates and palaeoenvironments from Charlie Quaternary Paleoecology in the Bonneville Basin. Lake Cave, northeast British Columbia. Canadian Utah Geological Survey, Bulletin 130, pp. 103- Journal of Earth Sciences, 25: 1545-1553. 121. Duke, G., Jegers, A., Leff, G., & Evanson, O. (1975). Broughton, J.M. (2000b). Terminal Pleistocene fish Gastric digestion in some raptors. Comparative remains from Homestead Cave, Utah, and Biochemistry and Physiology, 56A: 283-286. implications for fish biogeography in the England, A.S., Bechard, M.J. & Houston, C.S. (1997). Bonneville Basin. Copeia, 2000: 645-656. Swainson’s Hawk (Buteo swainsoni). In (A. Poole Broughton, J.M., Madsen, D.B. & Quade J. (2000). & F. Gill, Eds.) The Birds of North America, No. Fish remains from Homestead Cave and lake 265. The Academy of Natural Sciences, levels of the past 13,000 years in the Bonneville Philadelphia, and The American Ornithologists’ Basin. Quaternary Research, 53: 392-401. Union, Washington, D.C. Buehler, D.A. (2000). Bald Eagle (Haliaeetus Errington, P.L., Hammerstrom, F. & Hammerstrom, leucocephalus). In (A. Poole & F. Gill, Eds.) The F.N. (1940). The great horned owl and its prey in Birds of North America, No. 564. The Academy of north-central United States. Iowa Agricultural Natural Sciences, Philadelphia, and The American Experiment Station Research Bulletin, 277: 757- Ornithologists’ Union, Washington, D.C. 850.

93 Owl-deposited fish remains

Eshmeyer, W.N. & Herald, E.S. (1983). A Field Guide The Academy of Natural Sciences, Philadelphia, to Pacific Coast Fishes North America. Boston: and The American Ornithologists’ Union, Houghton Mifflin Co. Washington, D.C. Esteban, M., Castanet, J. & Sanchiz, B. (1998). Ille, R. (1991). Preference of prey size and profitability Inferring age and growth from remains of fossil in barn owls Tyto alba guttata. Behaviour, 116: and predated Recent anurans: a test case using 80-189. skeletochronology. Canadian Journal of Zoology, Janes, S.W. & Barss, J.M. (1985). Predation by three 76: 1689-1695. owl species on northern pocket gophers of Everitt, B.S. (1992). The Analysis of Contingency different body mass. Oecologia, 67: 76-81. Tables, Second Edition. London: Chapman & Hall. Johnsgard, P.A. (2002). North American Owls: Biology Follett, W.I. (1970). Fish remains from human and Natural History, Second Edition. Washington, coprolites and midden deposits obtained during D.C.: Smithsonian Institution Press. 1968 and 1969 at Lovelock Cave, Churchill Kimsey, J.B. (1954). The life history of the tui chub, County, Nevada. Contributions of the University of Siphateles bicolor (Girard), from Eagle Lake, California Archaeological Research Facility, 10: California. California Fish and Game, 40: 395- 163-175. 410. Follett, W.I. (1980). Fish remains from the Karlo Site Kochert, M.N., Steenhof, K., McIntyre, C.L. & Craig, (Ca-Las-7), Lassen County, California. Journal of E.H. (2002). Golden eagle (Aquila chrysaetos). In California and Great Basin Anthropology, 2: 114- (A. Poole & F. Gill, Eds.) The Birds of North 122. America, No. 684. The Academy of Natural Follett, W.I. (1982). An analysis of fish remains from Sciences, Philadelphia, and The American ten archaeological sites at Falcon Hill, Washoe Ornithologists’ Union, Washington, D.C. County, Nevada, with notes on fishing practices of König, C., Weick, F., & Becking, J.H. (1999). Owls: A the ethnographic Kuyuidikade Northern Paiute. Guide to the Owls of the World. New Haven: Yale Nevada State Museum Anthropological Papers, University Press. 18: 181-203. Kusmer, K.D. (1990). Taphonomy of owl pellet Gallup, F.N. (1949). Banding recoveries of Tyto alba. deposition. Journal of Paleontology, 64: 629-637. Bird-Banding, 20: 150. Lenton, G.M. (1984). The feeding and breeding Grayson, D.K. (1998). Moisture history and small ecology of Barn Owls Tyto alba in peninsular mammal community richness during the latest Malyasia. Ibis, 126: 551-575. Pleistocene and Holocene northern Bonneville Livingston, S.D. (2000). The Homestead Cave Basin, Utah. Quaternary Research, 49: 330-334. avifauna. In (D.B. Madsen, Ed.) Late Quaternary Grayson, D.K. (2000a). Mammalian responses to Paleoecology in the Bonneville Basin. Utah Middle Holocene climatic change in the Great Geological Survey, Bulletin 130, pp. 91-102. Basin of the western United States. Journal of Lyman, R.L. (1994). Vertebrate Taphonomy. Biogeography, 27: 181-192. Cambridge: Cambridge University Press. Grayson, D.K. (2000b). The Homestead Cave Lyman, R.L. & O’Brien, M.J. (1987). Plow-zone mammals. In (D. B. Madsen, Ed.) Late Quaternary zooarchaeology: fragmentation and identifiability. Paleoecology in the Bonneville Basin. Utah Journal of Field Archaeology, 14: 493-498. Geological Survey, Bulletin 130, pp. 67-89. Lyman, R.L., Power, E., & Lyman, R.J. (2003). Grayson, D.K. & Frey, C.J. (2004). Measuring skeletal Quantification and sampling of faunal remains in part representation in archaeological faunas. owl pellets. Journal of Taphonomy, 1: 3-14. Journal of Taphonomy, 2: 27-42. Madsen, D.B. (Ed.) (2000). Late Quaternary Hockett, B.S. (1995). Comparison of leporid bones in Paleoecology in the Bonneville Basin. Utah raptor pellets, raptor nests, and archaeological sites Geological Survey, Bulletin 130. in the Great Basin. North American Archaeologist, Madsen, D.B., Rhode, D., Grayson, D.K., Broughton, 16: 223-238. J.M., Livingston, S.D., Hunt, J., Quade, J., Hoffman, R. (1988). The contribution of raptorial birds Schmitt, D.N. & Shaver, M.W. (2001). Late to patterning in small mammal assemblages. Quaternary environmental change in the Paleobiology, 14: 81-90. Bonneville basin, western USA. Palaeogeography, Houston, C.S., Smith, D.G. & Rohner, C. (1998). Great Palaeoclimatology, Palaeoecology, 167: 243-271. Horned Owl (Bubo virginianus). In (A. Poole & F. Marti, C.D. (1974). Feeding ecology of four sympatric Gill, Eds.) The Birds of North America, No. 372. owls. The Condor, 76: 45-61.

94 Broughton et al.

Marti, C.D. (1990). Barn owl (Tyto alba). In (A. Poole comparison between South and North American, & F. Gill, Eds.) The Birds of North America, No. 1. barn owl, Tyto alba, populations. Journal of The Academy of Natural Sciences, Philadelphia, Archaeological Science, 25: 165-170. and The American Ornithologists’ Union, Schmitt, D.N. (2000). Faunal accumulations and Washington, D.C. animal foraging behavior in the Lakeside Mayhew, D.F. (1977). Avian predators as accumulators Mountains. In (D.B. Madsen, Ed.) Late of fossil mammal material. Boreas, 6: 25-31. Quaternary Paleoecology in the Bonneville Basin. Mollhagen, T.R., Wiley, R.W. & Packard, R.L. (1972). Utah Geological Survey, Bulletin 130, pp. 18-20. Prey remains in golden eagle nests: Texas and Schmitt, D.N. & Juell, K.E. (1994). Toward the New Mexico. Journal of Wildlife Management, 36: identification of coyote scatological faunal 784-792. accumulations in archaeological contexts. Journal Muñoz, M. & Casadevall, M. (1997). Fish remains of Archaeological Science, 21: 249-262. from Arbreda Cave (Serinyà, Girona), northeast Shaver, M.W. (2000). Archaeological materials from Spain, and their palaeoecological significance. Homestead Cave. In (D.B. Madsen, Ed.) Late Journal of Quaternary Science, 12: 111-115. Quaternary Paleoecology in the Bonneville Basin. Nagaoka, L. (2005). Differential recovery of Pacific Utah Geological Survey, Bulletin 130, pp. 59-66. Island fish remains. Journal of Archaeological Smith, G.R. (1985). Paleontology of Hidden Cave: Science, 32: 941-955. fish. In (D.H. Thomas, Ed). The Archaeology of Poole, A.F., Bierregaard, R.O. & Martell, M.S. (2002). Hidden Cave, Nevada. Anthropological Papers of Osprey (Pandion haliaetus). In (A. Poole & F. the American Museum of Natural History, 61: 171- Gill, Eds.) The Birds of North America, No. 683. 178. The Academy of Natural Sciences, Philadelphia, Stewart, K.M., LeBlanc, L., Matthiesen, D.P. & West, J. and The American Ornithologists’ Union, (1999). Microfaunal remains from a modern east Washington, D.C. African raptor roost: patterning and implications for Preston, C.R. & Beane, R.D. (1993). Red-tailed Hawk fossil bone scatters. Paleobiology, 25: 483-503. (Buteo jamaicensis). In (A. Poole & F. Gill, Eds.) Van Neer, W. & Morales Muñiz, A. (1992). “Fish The Birds of North America, No. 52. The Academy middens”: anthropogenic accumulations of fish of Natural Sciences, Philadelphia, and The remains and their bearing on archaeoichthyological American Ornithologists’ Union, Washington, analysis. Journal of Archaeological Science, 19: D.C. 683-695. Raczynski, J. & Ruprecht, A. (1974). The effect of Wheeler, A. & Jones, K.G. (1989). Fishes. digestion on the osteological composition of owl Cambridge: Cambridge University Press. pellets. Acta Ornithologica, 14: 25-38. Williams, J.E. & Bond, C.E. (1981). A new subspecies Rick, T.C., Erlandson, J.M. & Vellanoweth, R. (2001). of tui chub (Osteichthyes: Cyprinidae) from Guano Paleocoastal marine fishing on the Pacific Coast of Basin, Nevada and Oregon. The Southwestern the Americas: perspectives from Daisy Cave, Naturalist, 26: 223-230. California. American Antiquity, 66:595-613. Saavedra, B. & Simonetti, J.A. (1998). Small mammal taphonomy: intraspecific bone assemblage

Owl-deposited fish remains